Apolipoprotein E controls cerebrovascular integrity via cyclophilin A

Abstract

Human apolipoprotein E has three isoforms: APOE2, APOE3 and APOE4. APOE4 is a major genetic risk factor for Alzheimer's disease and is associated with Down's syndrome dementia and poor neurological outcome after traumatic brain injury and haemorrhage. Neurovascular dysfunction is present in normal APOE4 carriers and individuals with APOE4-associated disorders. In mice, lack of Apoe leads to blood-brain barrier (BBB) breakdown, whereas APOE4 increases BBB susceptibility to injury. How APOE genotype affects brain microcirculation remains elusive. Using different APOE transgenic mice, including mice with ablation and/or inhibition of cyclophilin A (CypA), here we show that expression of APOE4 and lack of murine Apoe, but not APOE2 and APOE3, leads to BBB breakdown by activating a proinflammatory CypA-nuclear factor-κB-matrix-metalloproteinase-9 pathway in pericytes. This, in turn, leads to neuronal uptake of multiple blood-derived neurotoxic proteins, and microvascular and cerebral blood flow reductions. We show that the vascular defects in Apoe-deficient and APOE4-expressing mice precede neuronal dysfunction and can initiate neurodegenerative changes. Astrocyte-secreted APOE3, but not APOE4, suppressed the CypA-nuclear factor-κB-matrix-metalloproteinase-9 pathway in pericytes through a lipoprotein receptor. Our data suggest that CypA is a key target for treating APOE4-mediated neurovascular injury and the resulting neuronal dysfunction and degeneration.

Figure 1. CypA deficiency or inhibition reverses BBB breakdown in Apoe^−/− and APOE4 mice

(a) Multiphoton microscopy of TMR-Dextran (white) in 6-month-old TR-APOE2, TR-APOE3, TR-APOE4, Apoe^−/− Ppia^+/+, Apoe^−/− Ppia^−/− and cyclosporine A-treated TR-APOE4 mice. Bar=20 mm. (b) CypA immunobloting in brain microvessels from apoE transgenic mice. (c) CypA (green) colocalization with PDGFRb-positive pericytes (red; yellow, merged) in hippocampal microvessels from Apoe^+/+, Apoe^−/− and GFAP-APOE4 mice. Blue, lectin-positive endothelium. Bar=10mm. (d) IgG neuronal uptake (green; lectin-positive vessels, blue) in Apoe^−/− Ppia^+/+, Apoe^−/− Ppia^−/−, TR-APOE4 Ppia^+/+ and TR-APOE4 Ppia^−/− mice. (e) Fibrin (red) and thrombin (green) in NeuN-positive neurons (blue) in the hippocampus of 9-month-old GFAP-APOE4 mice untreated and cyclosporine A-treated. a and c–e, representative results from 4–6 experiments. Scale bar, 10 μm.

Figure 2. CypA activates NF-kB-MMP-9 pathway causing BBB breakdown in Apoe^−/− and APOE4 mice

(a) Multiphoton microscopy of DQ-gelatin (green) in 8–9-month-old control, Apoe^−/− Ppia^+/+, Apoe^−/− Ppia^−/−, TR-APOE2, TR-APOE3, TR-APOE4 and cyclosporine A-treated Apoe^−/− Ppia^+/+ and TR-APOE4 mice. Red, cortical vessels. (b) Quantification of DQ-gelatin signal in apoE transgenic mice. Effects of cyclosporine A, PDTC, SB-3CT and CypA deletion in Apoe^−/− and TR-APOE4 mice. Mean±s.e.m., n=3–6 animals per group.(c) Gelatin zymography of brain tissue in control, Apoe^−/− Ppia^+/+, Apoe^−/− Ppia^−/−, TR-APOE3 and TR-APOE4 mice treated with vehicle, cyclosporine A, PDTC or SB-3CT. (d) MMP-9 (green) colocalization with CD13-positive pericytes (red; yellow, merged) in cortical microvessels from 9-month-old Apoe^+/+Ppia^+/+, Apoe^−/− Ppia^+/+ and TR-APOE4 mice. Blue, lectin-positive endothelium. Bar=10 mm. (e) Reduced collagen-IV, ZO-1, occludin and claudin-5 levels in 2-week-old Apoe^−/− and TR-APOE4 mice and reversal by CypA ablation (Apoe^−/− Ppia^+/+) and cyclosporine A (TR-APOE4). c and e, representative results from 4–6 experiments.

Figure 3. CypA ablation or inhibition reverses microvascular and CBF reductions in Apoe^−/− and APOE4 mice

(a) Capillary length in the hippocampus of apoE transgenic mice including Apoe^−/− Ppia^+/+, Apoe^−/− Ppia^−/− and GFAP-APOE4 and TR-APOE4 mice treated with cyclosporine A, SB-3CT or PDTC (mean±s.e.m., n=5 animals per group). (b) ¹⁴C-iodoantipyrine CBF autoradiograms in 9-month-old transgenic apoE mice. b, representative results from 6 experiments.

Figure 4. ApoE isoform-specific regulation of CypA-NF-kB-MMP-9 pathway in pericytes

(a) CypA mRNA quantification in Apoe^−/− pericytes after treatment with astrocyte-secreted apoE3, apoE4, siRNA silencing of LDL/apoE receptors and adenoviral-mediated re-expression of LRP1 minigene (Ad.m LRP1). Mean±s.e.m., n=3 independent cultures. (b) Proximity ligation imaging of apoE3 and apoE4 interaction with LRP1. (c–d) LRP1, CypA and MMP-9 immunodetection (c) and neuronal uptake (NeuN, green) of cadaverine-Alexa-Fluor-555 (red; yellow, merged) (d) in the hippocampus of 6-month-old GFAP-APOE3 mice after siLRP1 or control siRNA infusion. Blue, lectin-positive capillaries. (e) A schematic showing that astrocyte-secreted apoE3 and murine apoE, but not apoE4, signal to pericytes via LRP1 suppressing the CypA-NF-kB-MMP-9 pathway that causes BBB breakdown. b and c–d, representative results from 6 experiments.

Figure 5. Vascular defects in Apoe^−/− and APOE4 mice precede neuronal dysfunction

(a) The blood-brain barrier permeability surface (PS) product for tetramethylrhodamine (TMR)-dextran (40,000 Da) in the cortex and hippocampus of 2-week-old Apoe^+/+, Apoe^−/−, GFAP-APOE3 and GFAP-APOE4 mice measured by non-invasive fluorescence spectroscopy. (b) Representative time-lapse imaging profile analysis of fluorescent voltage sensitive dye (VSD) signal response in the hind-limb somatosensory cortex after stimulation in 2-week-old Apoe^+/+, Apoe^−/−, TR-APOE3 and TR-APOE4 mice. (c) VSD imaging of cortical responses to hind-limb stimulation in 4-month-old Apoe^+/+ and TR-APOE4 mice. (d) Representative VSD signal responses in the hind-limb somatosensory cortex region after stimulation in 4-month-old Apoe^+/+, Apoe^−/−, TR-APOE3 and TR-APOE4 mice. (e) Time to peak in fluorescent VSD signal after hind-limb stimulation in 4-month-old Apoe^+/+, Apoe^−/−, TR-APOE3, TR-APOE4, GFAP-APOE3 and GFAP-APOE4 mice and in Apoe^−/−, TR-APOE4 and GFAP-APOE4 mice treated with cyclosporine A, SB-3CT, PDTC or vehicle. a and e, mean±s.e.m., n= 5 animals per group.